Computing devices can be communicatively coupled to one another via a computer network to facilitate the transmission of data (e.g., network packets) between the computing devices. Data can be sent between the networked computing devices according to various protocols, for example using a “layered” networking model such as the open systems interconnection (OSI) model. In some instances, multiple communications links or paths may be utilized between computing devices for transmitting packets. However, due to congestion avoidance mechanisms, some methods of allocating packets among these links or paths may be inefficient or sub-optimal such as prioritizing packets according to connection latency. This may result in delays, loss of throughput, or unfairly penalizing some applications compared to others.